The overall objective of this project is to determine, with the aid of accurate models of the human breast, the most efficient and effective method of training the manual skills required for lump detection. Studies to date have substatially advanced the state of the art of breast simulation, resulting in models which simulate natural granularity and contain movable lumps of variable size and hardness. An automated training technology providing instruction and immediate feedback for approximation to the correct method of palpation will be completed and deployed in a field validation study involving 5,000 women. Data from this study will include records of lumps actually detected as a result of training provided. Concurrently, studies in progress will determine the proper frequency of home practice to maintain proficiency, the role of simple models in maintaining home practice, and the extent to which naturally occurring tissue change can serve a functional role in home practice.